Fast algorithms for composition and reversion of multivariate power series

39. R. P. Brent and H. T. Kung, Fast algorithms for composition and reversion of multivariate power series (preliminary version), in Proceedings of a Conference on Theoretical Computer Science Department of Computer Science, University of Waterloo, Waterloo, Ontario (August 1977), 149-158.

Abstract: dvi (3K), pdf (78K).

Extended abstract: pdf (112K).

Paper: pdf (631K).

Abstract

In our earlier papers [29, 45], we gave fast algorithms for manipulating dense univariate power series. In this paper, fast algorithms for composition and reversion of dense multivariate power series are presented. The new algorithms require substantially less operations than the best previously known algorithms. The relative advantage of the new algorithms increases with the number of variables in the multivariate power series.

Comments

Although called a "preliminary version" this was in fact the final version.

The univariate case is considered in Brent and Kung [45], and generalized composition is considered in Brent and Traub [50].

It should be noted that most multivariate power series occurring in practice are sparse, i.e. many of the coefficients are zero, but our algorithms do not take advantage of this.

Go to next publication

Return to Richard Brent's index page